The madness in the method

03 September 2008

Medical research without animals – can we get there? Dr John J Pippin, of the Physicians Committee for Responsible Medicine, considers the ethics from a different perspective.

No issue regarding medical research generates more dispute than the role of animal experimentation, and specifically the potential for its replacement with scientifically and ethically sound non-animal methods. Is the use of animals essential in medical research, or are there alternatives that should be adopted and further developed – in other words, is there a win-win answer that serves medical science and resolves the ethical issue?

Three important facts shed light on these questions. First, the association of animal experimentation with serious adverse physical and behavioural effects on animals has been demonstrated unequivocally – it is inherently cruel and harmful. Second, independently conducted opinion surveys demonstrate that public support for animal experimentation in the United States and the United Kingdom has declined over the last 50 years, and that the public would welcome the responsible replacement of animals in research.

Third, there is a growing consensus that medical research and testing can and should be done with scientifically sound non-animal methods. Some of the evidence follows.

Animal experimentation for drug and chemical testing
For decades, the default testing methods for the efficacy and safety of drugs and the toxicities of chemicals have relied heavily on the use of animals. Currently, however, government and non-government scientific organisations are strongly criticising the animal testing paradigm for its lack of reliability and poor scientific foundation.

Drug testing using animals is very inaccurate for predicting results in people. The US Food and Drug Administration (FDA) reports that the failure rate for drugs tested safely and effectively in preclinical studies (including animal tests) is 92%, and that this failure rate has increased from 86% in 1985, despite all efforts to improve animal modelling for drug testing.

About half of the 8% of drugs receiving FDA approval are later withdrawn or relabelled for serious or lethal adverse effects, and at least three-quarters of the remaining safe drugs are classified by FDA as 'me-too' drugs that provide 'little or no therapeutic gain' compared to currently available drugs. Further, more than 90% of approved drugs work in fewer than half of patients, and response rates (not cures) are as low as 25-30% for many drugs.

Thus, it takes about 100 drugs that are safe and effective in animal tests to produce just one unique and safe drug for humans, which then works only in a minority of patients and seldom provides a cure. There could hardly be stronger evidence that animal use for drug testing does not contribute to drug efficacy or safety.

The consequences of this failure are much worse than just the costs in time and money, as the current animal-based drug testing approach allows the approval and widespread use of dangerous drugs. The COX-2 inhibitor Vioxx was safe in at least eight studies in African green monkeys and five other animal species, but killed an estimated 60,000 Americans and 140,000 persons worldwide.

The monoclonal antibody TGN1412 was tested successfully on mice, rats, rabbits, and two species of monkeys. Yet TGN1412 rapidly caused critical illness in all six young men who received the drug in the first stage of human testing at London's Northwick Park Hospital in 2006. These six men nearly died, and all have permanent organ damage and future risk of cancers. The reaction to TGN1412 was the opposite of that seen in monkeys at up to 500 times higher doses, even though the key biological factors for this drug were identical for monkeys and humans¹.

The dramatic failures of drugs such as Vioxx and TGN1412 demonstrate that no level of certainty from animal testing can reliably predict drug effects in people. This truth led British immunotherapeutics expert David Glover to state: "The relevance of animal testing, whether artificially created disease models or healthy animals for toxicology, has to be very seriously questioned for testing of human specific biologic drugs²," and former Huntingdon Research Centre Scientific Director Ralph Heywood to state: "Toxicology is a science without a scientific underpinning³."

Moreover, the FDA acknowledged the inadequacy of animal testing for drug safety in its response to the September 2006 Institute of Medicine report 'The Future of Drug Safety: Promoting and Protecting the Health of the Public': 'The FDA is involved in an ongoing scientific collaboration intended to yield more sensitive, specific, and informative tests for drug organ toxicity than the toxicology screening techniques currently in use.'

The situation is quite similar regarding toxicity testing for chemicals. A July 2007 report from the National Research Council of the National Academies, commissioned by the US Environmental Protection Agency (EPA), made sweeping recommendations promoting replacement of animal testing with human specific methods. The report states that it should be possible to transform toxicity testing from a system based on whole animal testing to one based on methods specific for humans. In order to realise this goal, in January 2008, a collaborative agreement was announced by three key US science agencies – the EPA, the National Toxicology Program, and the National Institutes of Health Chemical Genomics Center. This memorandum of understanding is both a welcome acknowledgement and an important commitment from US federal agencies charged with protecting public health.

Animal experimentation for investigation of human diseases
Proponents of animal experimentation routinely make the sweeping claim that almost every major advance in human medicine during the last century has been due to animal research. Since this claim was first reported without supporting citation by the US Public Health Service in 1994, it has been recited regularly but has failed to hold up under scrutiny. Most recently, Robert Matthews' review in the Journal of the Royal Society of Medicine demonstrates that this claim is invalid, and further describes the inadequacies of predictive value and evidential weight from animal experimentation⁴.

Numerous peer reviewed scientific papers in the past five years demonstrate the unreliability of animal experimentation for predicting human clinical outcomes, and the suitability of non-animal methods to replace them. Given the failure of the animal experimentation model and the evidence that human specific methods are better models, persistence of many scientists' belief in the animal experimentation paradigm has been attributed to inertia, resistance to change, and conflict of interest – termed technological and institutional lock-in by American Economist Joshua Frank.

Unknown to most of the public, entire fields of medical discovery have produced little or nothing of value to humans from decades of animal experimentation. Although more than 80 HIV/AIDS vaccines have been successful in non-human primate studies, as of June 2008, every one of 63 preventive vaccine trials and 32 therapeutic vaccine trials reported in the US federal clinical trials database has failed to demonstrate benefit to humans. Similarly, every one of more than 150 successful animal stroke treatments tested in humans has failed.

Every one of at least two dozen animal diabetes cures has failed in humans, and the traditional mouse diabetes model has been discredited by scientists. Every one of more than 20 human trials in patients suffering from spinal cord injuries has failed to confirm benefits seen in animal studies. Decades of animal experimentation have failed to cure or substantially improve dozens of chronic diseases, and the Director of the United States' national war on cancer reported in 1997 that no substantial progress was made after a quarter century of effort focused on animal modelled drug development.

The traditional mouse models for cancer have been widely discredited by researchers, based on failure to correlate with human cancers or to predict treatment effects. The US National Cancer Institute (NCI) developed the DTP Human Tumor Cell Line Screen, a panel of 60 human tumour cell lines, to replace unreliable animal testing for identification of effective cancer treatments. According to former NCI Director Dr Richard Klausner, "we have cured mice of cancer for decades, and it simply didn't work in humans⁵."

Finally, the renaissance in medical science promised from the use of genetically modified (GM) animals, predominantly mice, has not occurred. To the contrary, it has been demonstrated that the reported links between genes and diseases are often not valid, that factors other than genes can invalidate the results produced in GM animals, and that identical genes often function differently in mice and humans – undermining the very premises on which GM animal science is based.

Where do we go from here?
There are no perfect tools to investigate human diseases or to test drugs and chemicals. But it is clear that using other species for these purposes is a failure in scientific and practical terms, and thus is also especially egregious in ethical terms. When scientific studies prove that identical receptor biology produces safe results in monkeys and deadly consequences in humans, that genetically identical rats produce different drug testing results, and that identical human twins differ in important disease and drug response related measures, how can we continue to believe that answers will come from studying non-human animals?

It is often said that animal experiments must continue until 'better' methods are available. There are such methods, including, but not limited to, such technologies as computational science, bioinformatics, systems biology, in vitro techniques, tissue engineering, microfluidics, stem cell methods, human tissue studies, genetic methods, and microdosing. More importantly, even better methods will be found if the emphasis shifts from tweaking animal models to developing human specific methods.

So consider the ethics of animal experimentation from a different perspective. Is it not unethical toward people to persist in the use of animal-based research and testing methods when the results are so poor, the consequences are so dire, and the answers are at hand?

Cardiologist John J Pippin, MD, is a senior medical and research adviser for the Physicians Committee for Responsible Medicine, a Washington, D.C.-based non-profit organisation that promotes preventive medicine, conducts clinical research, and encourages higher standards for ethics and effectiveness in research.

1 Kenter M J H & Cohen A F (2006) 'Establishing risk of human experimentation with drugs: lessons from TGN1412', The Lancet, 368, 1387-91
2 Mitchell (2007) 'Critics pan timid European response to TeGenero disaster', Nat Biotechnol, 25, 485-6
3 Allen A (2006) 'Of mice or men: the problems with animal testing', Slate, 1st June, accessed 12th June 2008: http://www.slate.com/id/2142814
4 Matthews R A J (2008) 'Medical progress depends on animal models – doesn't it?' J R Soc Med, 101, 95-8
5 Cimons M, Getlin J & Maugh T H III (1998) 'Cancer Drugs Face Long Road From Mice to Men; Medicine: Doctors downplay excitement over report. Questions raised about how media handle such advances', Los Angeles Times, 6th May, page A1